【作者】 郭龙飞；

【导师】 吕廷杰；

【作者基本信息】 北京邮电大学 ， 管理科学与工程， 2013， 博士

【摘要】 随着互联网、移动互联网愈发深刻地影响着人们的社会行为,用户暴露在网络环境中的维度和几率大大增加,隐私威胁已经成为关乎每个网络用户切身利益的重要问题。特别是在功能强大的智能终端、丰富多彩的服务应用、日益增长的网络关联等因素作用下,用户越来越多的隐私不可避免地主动、无意或被恶意泄露。用户需要将控制隐私的权利把握在自己手中,网络应用提供商则希望利用尽可能多的隐私信息获取更多的利益,其他利益相关方也想从中赚取更多价值。可见,隐私问题的深入研究具有重大的现实意义。由于隐私概念具有人类社会属性,故隐私的研究可以依托社会网络,而社交网络的出现为隐私问题提供了很好的研究环境,它不仅拥有与现实社会极为接近的用户规模、相互关系、复杂结构、行为规律,还拥有现实社会无法相比的易获取、易分析、易验证等特征。因此,本文以社交网络为研究对象,聚焦于用户隐私关注产生影响的关键因素以及由此引发的用户行为规律等相关问题,具体研究内容包括以下三部分：(1)社交网络用户隐私关注动态影响因素研究。以往对于用户隐私关注影响因素的研究,大都认为影响因素是静态的,但随着网络将人们的社交活动数量加大、频率提高,原本变化缓慢的隐私感知开始更容易被用户察觉到,特别是更密集、更重大的隐私负面事件直接或间接地被用户接收到时,其对隐私威胁的判断势必发生变化。基于此,本研究站在动态的视角,结合以往理论基础和当前研究热点,建立隐私关注动态影响模型。考虑用户所受直接和间接影响、不同类型社交网站的不同作用特征,发现与隐私相关的突发事件和媒体宣传对用户的隐私关注有着非常明显的影响。在对比情感性、工具性、混合性、无连带性社交网站后,发现各类社交网站中,用户隐私关注受动态因素影响的结果有很大不同,其中新浪微博为代表的混合性社交网站呈现出最贴近假设、最贴近用户社会行为特征的结果。因此,后续隐私行为规律研究选取了新浪微博平台作为实证研究对象。(2)社交网络利益驱动下的隐私行为演化博弈研究。以往对于隐私悖论的探讨,大都集中在局中人当下利益的博弈,但社交网络环境下,人们具有长期的社会行为,博弈局中人面对利益时采取的策略更多是长期策略。因此,本研究采用演化博弈思想,将用户和网站双方看成两个群体,基于前部分用户利益感知对其行为的影响以及以往研究对利益成本和收益的描述,通过搭建收益矩阵和群体复制者动态方程,得到隐私利益博弈的演化博弈模型。进而,引入突发事件干扰因素后,调整演化博弈模型,获得不同条件下的演化结果,结合仿真后的结果对比图,清晰地观察到演化规律。最后,通过对比用户方与网站方期望结果,讨论出社会意义最大的隐私利益博弈的演化规律,为网站方、利益相关方、媒体、监管方等角色提出建议,以期保护用户利益,并最大化社会效益。后续研究将此部分结论放到真实社交网络环境中进行观察,所得结果更有实践指导意义。此外,基于此部分宏观规律的指导,后续部分深入分析了微观层面用户隐私行为演化规律。(3)社交网络动态隐私关注条件下的用户行为规律研究。以往复杂网络中用户行为规律研究方法及思路已经比较成熟,但与隐私相关的用户行为规律研究还不多见。近年来,一些具有隐私属性的网络应用快速发展,使得新环境下的隐私研究更为迫切。本研究聚焦于LBS典型隐私信息应用,借助与新浪微博平台的合作,通过社会网络分析方法及大数据处理软件对大量隐私信息进行深入分析,从而获得用户隐私行为与业务采纳扩散规律以及突发事件对隐私行为影响规律。最后,结合以上各部分研究结果,以社会网络中隐私相关行为规律为依据,对社交网络发展、隐私业务应用、隐私监管策略、用户隐私安全策略等问题提出相应建议。本研究的创新之处有三点：(1)构建了社交网络隐私关注动态影响因素模型。以突发事件和媒体曝光为考察对象,在动态环境中观察用户隐私关注的变化,对比个人和群体对隐私关注的不同作用效果,讨论不同功能社交网络对隐私关注的不同作用结果,发现用户社交网络隐私关注的影响规律；(2)构建了社交网络突发事件影响下隐私行为演化博弈模型。考虑在突发事件扰动下,用户群体和网站群体在不同利益策略下,用户的选择和突变行为演化规律,以此讨论用户与网站以及第三方策略对社交网络发展所起作用；(3)探究了社交网络隐私业务采纳行为规律。通过新浪微博最新、最典型的隐私属性业务LBS应用,获取研究数据,利用社会网络分析方法,发现隐私业务采纳规律受隐私相关突发事件的影响。更多还原

【Abstract】 Along with the development of the Internet and mobile Internet, human’s social behaviors have been deeply influenced. People are more easily to be exposed to a net environment. Therefore, privacy threat has become been a very crucial issue which is related to the interest of every consumer consumers. Gradually, an even increasing amount of Privacy has been unavoidablyrevealed by accident or by malice, especially under the effect of intelligent terminal, service application and network association. Users want to own the right of privacy controlling, while the internet application providers wish for users’privacy in order to gain more profit, the same as the other interested parties. As we can see, it is significant to do deep researches in the area of privacy.The concept of privacy has the attribute of human society, so the research of privacy can be rely on social network. The appearance of social network provides a very good environment for privacy researching. It not only has the scale of users, interrelation, complex construction, behavior rules which are nearly the same as the real society, but also owns the characteristics of accessibility, analyzability and verifiability which cannot be get in the real society. Thus, taken social network as the research object, this paper focuses on the factors that influence privacy concerns and the behavior rules which are related to privacy concerns. The specific research contents are as follows:(i)Research on Dynamic Factors that Influence Privacy Concern of-Social Network.Researches on factors that influence privacy concern before are mainly based on static factors, but with the increase of the number and frequency of social activities, privacy awareness is more likely to be aware. Particularly when some great negative events are faced by users, they will strengthen the judgment of privacy threats. Based on-this background, this paper combines classic theory with new research hotspots and sets up a model of factors that influence privacy concern from the view of dynamic. Taking the direct and indirect influence, different features in different kinds of social networks into consideration, the author found out that emergency and media exposure do great impact on users’privacy concerns. After comparing different kinds of social networks (emotional, instrumental, uncorrelated, mixed), the author findsfigured out that the results are quite different toward different websites. Among them, Sina Weibo, which represents mixed websites, obtains the best result in users’characteristics of social activity. So the following research on behavior rules will rely on the platform of Sina Weibo.(ii)Evolutionary Game Research on Privacy Behaviors based on Profit-Driven in Social Network.Researches on privacy paradox before mainly focus on immediate interest, but in the environment of social network, people’s pattern of social behaviors is likely to last longer and strategies are mostly in long-term. Therefore, this paper proposes an evolutionary game which contains users and website service providers. Consulting the above results that explained the users’interest perception and its effect on behavior and former researches on cost and profit, the author sets up a privacy evolution game model through constructing a pay-off matrix and group replication dynamic equation. Moreover, the author adds the factor of emergency and adjusts the game model, further gets results that are in different conditions. We can clearly obtain the evolutionary rules by combining with the simulation experiment. At last, through comparing the results in different groups, we pointed out the significant evolutionary rules in privacy interests which can be good advices to website service providers, stakeholders, media and supervisors, in order to protect the interests of consumers and maximize social benefits. Follow-up study will test the conclusion which is drawn in this part by observing in the real social network environment. The results will be more significant in practicalness. In addition, the next part will deeply analyze the evolutionary rules of users’privacy behaviors in micro-level, based on macro-level rules from this part.(iii)Research on Users’Behavior Rules in the Condition of Dynamic Privacy Concern in Social Network.There have been a lot of researches on users’behavior rules in complex network. The research methods are already mature, but users’ behavior rules with privacy are immature now. In recent years, privacy researches are imperative because of the rapid development of web applications which may cause privacy disclosure. This research mainly focuses on classic LBS applications. Via the cooperation with Sina Weibo, we obtain the rules of privacy behaviors and business adoption diffusion, as well as the effect rules of how emergencies affect privacy behaviors by analyzing a lot of privacy information through social network analysis methods and data processing softwares. Finally, by summarizing the above research results, the author proposes some suggestions on development of social network, privacy business application, strategies on privacy supervision, strategies on privacy safety etc. according to the privacy behavior rules in social network.The innovation points in this research are as follows.(1) We set a model of dynamic factors that influence privacy concerns in social network. We select emergency and media exposure to be main factors, and observe the change of users’privacy concern in a dynamic environment, as well as compare the influence extent of different factors from both personal view and group view. Then we discuss different results from different social networks and find out the effect rules of users’ privacy concerns in social network;(2) An evolutionary game model under the influence of emergencies in social network is set. Considering that under the disturbance of emergency, the users’group and the website service providers’group will choose different interest strategies, so that we can gain an evolutionary rule of users’choices and behavioral mutation. On this account, we can discuss how different strategies from diverse groups will promote the development of social network;(3) explore behavior rules of privacy business adoption in social network. We gain data from the latest and the most classic LBS application which featured in privacy attribute through Sina Weibo. Through analysis method in social network, we find out privacy-related emergencies will affect the privacy business adoption rules.更多还原